A method and system for simultaneous quantitative multiparametric MRI are provided in the disclosure. The method includes: designing a magnetic resonance imaging (MRI) sequence which includes a signal excitation part, a shift gradient part, and a data acquisition part; generating training samples for a deep neural network according to the MRI sequence; training the deep neural network by using the training samples, to obtain a trained deep neural network; and reconstructing multiple magnetic-resonance parametric maps by using the trained deep neural network and k-space data acquired by the MRI sequence. The disclosure can implement quantitative multiparametric MRI, and can correct image distortions caused by magnetic field inhomogeneity.